BSD 4_3_Net_2 release

[unix-history] / usr / src / sys / kern / kern_descrip.c

diff --git a/usr/src/sys/kern/kern_descrip.c b/usr/src/sys/kern/kern_descrip.c
index 3828ce7..26c8c54 100644 (file)
--- a/usr/src/sys/kern/kern_descrip.c
+++ b/usr/src/sys/kern/kern_descrip.c
@@ -1,28 +1,41 @@
 /*
 /*

- * Copyright (c) 1982, 1986, 1989 Regents of the University of California.
+ * Copyright (c) 1982, 1986, 1989, 1991 Regents of the University of California.

  * All rights reserved.
  *
  * All rights reserved.
  *

- * Redistribution is only permitted until one year after the first shipment
- * of 4.4BSD by the Regents.  Otherwise, redistribution and use in source and
- * binary forms are permitted provided that: (1) source distributions retain
- * this entire copyright notice and comment, and (2) distributions including
- * binaries display the following acknowledgement:  This product includes
- * software developed by the University of California, Berkeley and its
- * contributors'' in the documentation or other materials provided with the
- * distribution and in all advertising materials mentioning features or use
- * of this software.  Neither the name of the University nor the names of
- * its contributors may be used to endorse or promote products derived from
- * this software without specific prior written permission.
- * THIS SOFTWARE IS PROVIDED AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
- * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *     This product includes software developed by the University of
+ *     California, Berkeley and its contributors.
+ * 4. Neither the name of the University nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.

  *
  *

- *     @(#)kern_descrip.c      7.16 (Berkeley) 6/28/90
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ *     @(#)kern_descrip.c      7.28 (Berkeley) 6/25/91

  */
 
 #include "param.h"
 #include "systm.h"
  */
 
 #include "param.h"
 #include "systm.h"

-#include "user.h"
+#include "filedesc.h"

 #include "kernel.h"
 #include "vnode.h"
 #include "proc.h"
 #include "kernel.h"
 #include "vnode.h"
 #include "proc.h"


@@ -31,10 +44,16 @@
 #include "socketvar.h"
 #include "stat.h"
 #include "ioctl.h"
 #include "socketvar.h"
 #include "stat.h"
 #include "ioctl.h"

+#include "fcntl.h"
+#include "malloc.h"
+#include "syslog.h"
+#include "resourcevar.h"

 
 /*
  * Descriptor management.
  */
 
 /*
  * Descriptor management.
  */

+struct file *filehead; /* head of list of open files */
+int nfiles;            /* actual number of open files */

 
 /*
  * System calls on descriptors.
 
 /*
  * System calls on descriptors.


@@ -46,7 +65,7 @@ getdtablesize(p, uap, retval)
        int *retval;
 {
 
        int *retval;
 {
 

-       *retval = NOFILE;
+       *retval = p->p_rlimit[RLIMIT_OFILE].rlim_cur;

        return (0);
 }
 
        return (0);
 }
 


@@ -61,6 +80,7 @@ dup(p, uap, retval)
        } *uap;
        int *retval;
 {
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        struct file *fp;
        int fd, error;
 
        struct file *fp;
        int fd, error;
 


@@ -69,15 +89,16 @@ dup(p, uap, retval)
         */
        if (uap->i &~ 077) { uap->i &= 077; return (dup2(p, uap, retval)); }
 
         */
        if (uap->i &~ 077) { uap->i &= 077; return (dup2(p, uap, retval)); }
 

-       if ((unsigned)uap->i >= NOFILE || (fp = u.u_ofile[uap->i]) == NULL)
+       if ((unsigned)uap->i >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[uap->i]) == NULL)

                return (EBADF);
                return (EBADF);

-       if (error = ufalloc(0, &fd))
+       if (error = fdalloc(p, 0, &fd))

                return (error);
                return (error);

-       u.u_ofile[fd] = fp;
-       u.u_pofile[fd] = u.u_pofile[uap->i] &~ UF_EXCLOSE;
+       fdp->fd_ofiles[fd] = fp;
+       fdp->fd_ofileflags[fd] = fdp->fd_ofileflags[uap->i] &~ UF_EXCLOSE;

        fp->f_count++;
        fp->f_count++;

-       if (fd > u.u_lastfile)
-               u.u_lastfile = fd;
+       if (fd > fdp->fd_lastfile)
+               fdp->fd_lastfile = fd;

        *retval = fd;
        return (0);
 }
        *retval = fd;
        return (0);
 }


@@ -88,37 +109,43 @@ dup(p, uap, retval)
 /* ARGSUSED */
 dup2(p, uap, retval)
        struct proc *p;
 /* ARGSUSED */
 dup2(p, uap, retval)
        struct proc *p;

-       register struct args {
-               int     i;
-               int     j;
+       struct args {
+               u_int   from;
+               u_int   to;

        } *uap;
        int *retval;
 {
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        register struct file *fp;
        register struct file *fp;

-       int error;
+       register u_int old = uap->from, new = uap->to;
+       int i, error;

 
 

-       if ((unsigned)uap->i >= NOFILE || (fp = u.u_ofile[uap->i]) == NULL)
-               return (EBADF);
-       if (uap->j < 0 || uap->j >= NOFILE)
+       if (old >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[old]) == NULL ||
+           new >= p->p_rlimit[RLIMIT_OFILE].rlim_cur)

                return (EBADF);
                return (EBADF);

-       *retval = uap->j;
-       if (uap->i == uap->j)
+       *retval = new;
+       if (old == new)

                return (0);
                return (0);

-       if (u.u_ofile[uap->j]) {
-               if (u.u_pofile[uap->j] & UF_MAPPED)
-                       munmapfd(uap->j);
-               error = closef(u.u_ofile[uap->j]);
+       if (new >= fdp->fd_nfiles) {
+               if (error = fdalloc(p, new, &i))
+                       return (error);
+               if (new != i)
+                       panic("dup2: fdalloc");
+       } else if (fdp->fd_ofiles[new]) {
+               if (fdp->fd_ofileflags[new] & UF_MAPPED)
+                       (void) munmapfd(p, new);
+               /*
+                * dup2() must succeed even if the close has an error.
+                */
+               (void) closef(fdp->fd_ofiles[new], p);

        }
        }

-       u.u_ofile[uap->j] = fp;
-       u.u_pofile[uap->j] = u.u_pofile[uap->i] &~ UF_EXCLOSE;
+       fdp->fd_ofiles[new] = fp;
+       fdp->fd_ofileflags[new] = fdp->fd_ofileflags[old] &~ UF_EXCLOSE;

        fp->f_count++;
        fp->f_count++;

-       if (uap->j > u.u_lastfile)
-               u.u_lastfile = uap->j;
-       /*
-        * dup2() must suceed even though the close had an error.
-        */
-       error = 0;              /* XXX */
-       return (error);
+       if (new > fdp->fd_lastfile)
+               fdp->fd_lastfile = new;
+       return (0);

 }
 
 /*
 }
 
 /*


@@ -128,31 +155,34 @@ dup2(p, uap, retval)
 fcntl(p, uap, retval)
        struct proc *p;
        register struct args {
 fcntl(p, uap, retval)
        struct proc *p;
        register struct args {

-               int     fdes;
+               int     fd;

                int     cmd;
                int     arg;
        } *uap;
        int *retval;
 {
                int     cmd;
                int     arg;
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        register struct file *fp;
        register char *pop;
        register struct file *fp;
        register char *pop;

-       int i, error;
+       struct vnode *vp;
+       int i, tmp, error, flg = F_POSIX;
+       struct flock fl;

 
 

-       if ((unsigned)uap->fdes >= NOFILE ||
-           (fp = u.u_ofile[uap->fdes]) == NULL)
+       if ((unsigned)uap->fd >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[uap->fd]) == NULL)

                return (EBADF);
                return (EBADF);

-       pop = &u.u_pofile[uap->fdes];
+       pop = &fdp->fd_ofileflags[uap->fd];

        switch(uap->cmd) {
        case F_DUPFD:
        switch(uap->cmd) {
        case F_DUPFD:

-               if (uap->arg < 0 || uap->arg >= NOFILE)
+               if ((unsigned)uap->arg >= p->p_rlimit[RLIMIT_OFILE].rlim_cur)

                        return (EINVAL);
                        return (EINVAL);

-               if (error = ufalloc(uap->arg, &i))
+               if (error = fdalloc(p, uap->arg, &i))

                        return (error);
                        return (error);

-               u.u_ofile[i] = fp;
-               u.u_pofile[i] = *pop &~ UF_EXCLOSE;
+               fdp->fd_ofiles[i] = fp;
+               fdp->fd_ofileflags[i] = *pop &~ UF_EXCLOSE;

                fp->f_count++;
                fp->f_count++;

-               if (i > u.u_lastfile)
-                       u.u_lastfile = i;
+               if (i > fdp->fd_lastfile)
+                       fdp->fd_lastfile = i;

                *retval = i;
                return (0);
 
                *retval = i;
                return (0);
 


@@ -165,23 +195,100 @@ fcntl(p, uap, retval)
                return (0);
 
        case F_GETFL:
                return (0);
 
        case F_GETFL:

-               *retval = fp->f_flag + FOPEN;
+               *retval = OFLAGS(fp->f_flag);

                return (0);
 
        case F_SETFL:
                return (0);
 
        case F_SETFL:

-               fp->f_flag &= FCNTLCANT;
-               fp->f_flag |= (uap->arg-FOPEN) &~ FCNTLCANT;
-               if (error = fset(fp, FNDELAY, fp->f_flag & FNDELAY))
+               fp->f_flag &= ~FCNTLFLAGS;
+               fp->f_flag |= FFLAGS(uap->arg) & FCNTLFLAGS;
+               tmp = fp->f_flag & FNONBLOCK;
+               error = (*fp->f_ops->fo_ioctl)(fp, FIONBIO, (caddr_t)&tmp, p);
+               if (error)

                        return (error);
                        return (error);

-               if (error = fset(fp, FASYNC, fp->f_flag & FASYNC))
-                       (void) fset(fp, FNDELAY, 0);
+               tmp = fp->f_flag & FASYNC;
+               error = (*fp->f_ops->fo_ioctl)(fp, FIOASYNC, (caddr_t)&tmp, p);
+               if (!error)
+                       return (0);
+               fp->f_flag &= ~FNONBLOCK;
+               tmp = 0;
+               (void) (*fp->f_ops->fo_ioctl)(fp, FIONBIO, (caddr_t)&tmp, p);

                return (error);
 
        case F_GETOWN:
                return (error);
 
        case F_GETOWN:

-               return (fgetown(fp, retval));
+               if (fp->f_type == DTYPE_SOCKET) {
+                       *retval = ((struct socket *)fp->f_data)->so_pgid;
+                       return (0);
+               }
+               error = (*fp->f_ops->fo_ioctl)
+                       (fp, (int)TIOCGPGRP, (caddr_t)retval, p);
+               *retval = -*retval;
+               return (error);

 
        case F_SETOWN:
 
        case F_SETOWN:

-               return (fsetown(fp, uap->arg));
+               if (fp->f_type == DTYPE_SOCKET) {
+                       ((struct socket *)fp->f_data)->so_pgid = uap->arg;
+                       return (0);
+               }
+               if (uap->arg <= 0) {
+                       uap->arg = -uap->arg;
+               } else {
+                       struct proc *p1 = pfind(uap->arg);
+                       if (p1 == 0)
+                               return (ESRCH);
+                       uap->arg = p1->p_pgrp->pg_id;
+               }
+               return ((*fp->f_ops->fo_ioctl)
+                       (fp, (int)TIOCSPGRP, (caddr_t)&uap->arg, p));
+
+       case F_SETLKW:
+               flg |= F_WAIT;
+               /* Fall into F_SETLK */
+
+       case F_SETLK:
+               if (fp->f_type != DTYPE_VNODE)
+                       return (EBADF);
+               vp = (struct vnode *)fp->f_data;
+               /* Copy in the lock structure */
+               error = copyin((caddr_t)uap->arg, (caddr_t)&fl, sizeof (fl));
+               if (error)
+                       return (error);
+               if (fl.l_whence == SEEK_CUR)
+                       fl.l_start += fp->f_offset;
+               switch (fl.l_type) {
+
+               case F_RDLCK:
+                       if ((fp->f_flag & FREAD) == 0)
+                               return (EBADF);
+                       p->p_flag |= SADVLCK;
+                       return (VOP_ADVLOCK(vp, (caddr_t)p, F_SETLK, &fl, flg));
+
+               case F_WRLCK:
+                       if ((fp->f_flag & FWRITE) == 0)
+                               return (EBADF);
+                       p->p_flag |= SADVLCK;
+                       return (VOP_ADVLOCK(vp, (caddr_t)p, F_SETLK, &fl, flg));
+
+               case F_UNLCK:
+                       return (VOP_ADVLOCK(vp, (caddr_t)p, F_UNLCK, &fl,
+                               F_POSIX));
+
+               default:
+                       return (EINVAL);
+               }
+
+       case F_GETLK:
+               if (fp->f_type != DTYPE_VNODE)
+                       return (EBADF);
+               vp = (struct vnode *)fp->f_data;
+               /* Copy in the lock structure */
+               error = copyin((caddr_t)uap->arg, (caddr_t)&fl, sizeof (fl));
+               if (error)
+                       return (error);
+               if (fl.l_whence == SEEK_CUR)
+                       fl.l_start += fp->f_offset;
+               if (error = VOP_ADVLOCK(vp, (caddr_t)p, F_GETLK, &fl, F_POSIX))
+                       return (error);
+               return (copyout((caddr_t)&fl, (caddr_t)uap->arg, sizeof (fl)));

 
        default:
                return (EINVAL);
 
        default:
                return (EINVAL);


@@ -189,66 +296,6 @@ fcntl(p, uap, retval)
        /* NOTREACHED */
 }
 
        /* NOTREACHED */
 }
 

-fset(fp, bit, value)
-       struct file *fp;
-       int bit, value;
-{
-
-       if (value)
-               fp->f_flag |= bit;
-       else
-               fp->f_flag &= ~bit;
-       return (fioctl(fp, (int)(bit == FNDELAY ? FIONBIO : FIOASYNC),
-           (caddr_t)&value));
-}
-
-fgetown(fp, valuep)
-       struct file *fp;
-       int *valuep;
-{
-       int error;
-
-       switch (fp->f_type) {
-
-       case DTYPE_SOCKET:
-               *valuep = ((struct socket *)fp->f_data)->so_pgid;
-               return (0);
-
-       default:
-               error = fioctl(fp, (int)TIOCGPGRP, (caddr_t)valuep);
-               *valuep = -*valuep;
-               return (error);
-       }
-}
-
-fsetown(fp, value)
-       struct file *fp;
-       int value;
-{
-
-       if (fp->f_type == DTYPE_SOCKET) {
-               ((struct socket *)fp->f_data)->so_pgid = value;
-               return (0);
-       }
-       if (value > 0) {
-               struct proc *p = pfind(value);
-               if (p == 0)
-                       return (ESRCH);
-               value = p->p_pgrp->pg_id;
-       } else
-               value = -value;
-       return (fioctl(fp, (int)TIOCSPGRP, (caddr_t)&value));
-}
-
-fioctl(fp, cmd, value)
-       struct file *fp;
-       int cmd;
-       caddr_t value;
-{
-
-       return ((*fp->f_ops->fo_ioctl)(fp, cmd, value));
-}
-

 /*
  * Close a file descriptor.
  */
 /*
  * Close a file descriptor.
  */


@@ -256,24 +303,28 @@ fioctl(fp, cmd, value)
 close(p, uap, retval)
        struct proc *p;
        struct args {
 close(p, uap, retval)
        struct proc *p;
        struct args {

-               int     fdes;
+               int     fd;

        } *uap;
        int *retval;
 {
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        register struct file *fp;
        register struct file *fp;

+       register int fd = uap->fd;

        register u_char *pf;
 
        register u_char *pf;
 

-       if ((unsigned)uap->fdes >= NOFILE ||
-           (fp = u.u_ofile[uap->fdes]) == NULL)
+       if ((unsigned)fd >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[fd]) == NULL)

                return (EBADF);
                return (EBADF);

-       pf = (u_char *)&u.u_pofile[uap->fdes];
+       pf = (u_char *)&fdp->fd_ofileflags[fd];

        if (*pf & UF_MAPPED)
        if (*pf & UF_MAPPED)

-               munmapfd(uap->fdes);
-       u.u_ofile[uap->fdes] = NULL;
-       while (u.u_lastfile >= 0 && u.u_ofile[u.u_lastfile] == NULL)
-               u.u_lastfile--;
+               (void) munmapfd(p, fd);
+       fdp->fd_ofiles[fd] = NULL;
+       while (fdp->fd_lastfile > 0 && fdp->fd_ofiles[fdp->fd_lastfile] == NULL)
+               fdp->fd_lastfile--;
+       if (fd < fdp->fd_freefile)
+               fdp->fd_freefile = fd;

        *pf = 0;
        *pf = 0;

-       return (closef(fp));
+       return (closef(fp, p));

 }
 
 /*
 }
 
 /*


@@ -283,22 +334,23 @@ close(p, uap, retval)
 fstat(p, uap, retval)
        struct proc *p;
        register struct args {
 fstat(p, uap, retval)
        struct proc *p;
        register struct args {

-               int     fdes;
+               int     fd;

                struct  stat *sb;
        } *uap;
        int *retval;
 {
                struct  stat *sb;
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        register struct file *fp;
        struct stat ub;
        int error;
 
        register struct file *fp;
        struct stat ub;
        int error;
 

-       if ((unsigned)uap->fdes >= NOFILE ||
-           (fp = u.u_ofile[uap->fdes]) == NULL)
+       if ((unsigned)uap->fd >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[uap->fd]) == NULL)

                return (EBADF);
        switch (fp->f_type) {
 
        case DTYPE_VNODE:
                return (EBADF);
        switch (fp->f_type) {
 
        case DTYPE_VNODE:

-               error = vn_stat((struct vnode *)fp->f_data, &ub);
+               error = vn_stat((struct vnode *)fp->f_data, &ub, p);

                break;
 
        case DTYPE_SOCKET:
                break;
 
        case DTYPE_SOCKET:


@@ -315,72 +367,137 @@ fstat(p, uap, retval)
 }
 
 /*
 }
 
 /*

- * Allocate a user file descriptor.
+ * Allocate a file descriptor for the process.

  */
  */

-ufalloc(want, result)
-       register int want;
+int fdexpand;
+
+fdalloc(p, want, result)
+       struct proc *p;
+       int want;

        int *result;
 {
        int *result;
 {

+       register struct filedesc *fdp = p->p_fd;
+       register int i;
+       int lim, last, nfiles;
+       struct file **newofile;
+       char *newofileflags;

 
 

-       for (; want < NOFILE; want++) {
-               if (u.u_ofile[want] == NULL) {
-                       u.u_pofile[want] = 0;
-                       if (want > u.u_lastfile)
-                               u.u_lastfile = want;
-                       *result = want;
-                       return (0);
+       /*
+        * Search for a free descriptor starting at the higher
+        * of want or fd_freefile.  If that fails, consider
+        * expanding the ofile array.
+        */
+       lim = p->p_rlimit[RLIMIT_OFILE].rlim_cur;
+       for (;;) {
+               last = min(fdp->fd_nfiles, lim);
+               if ((i = want) < fdp->fd_freefile)
+                       i = fdp->fd_freefile;
+               for (; i < last; i++) {
+                       if (fdp->fd_ofiles[i] == NULL) {
+                               fdp->fd_ofileflags[i] = 0;
+                               if (i > fdp->fd_lastfile)
+                                       fdp->fd_lastfile = i;
+                               if (want <= fdp->fd_freefile)
+                                       fdp->fd_freefile = i;
+                               *result = i;
+                               return (0);
+                       }

                }
                }

+
+               /*
+                * No space in current array.  Expand?
+                */
+               if (fdp->fd_nfiles >= lim)
+                       return (EMFILE);
+               if (fdp->fd_nfiles < NDEXTENT)
+                       nfiles = NDEXTENT;
+               else
+                       nfiles = 2 * fdp->fd_nfiles;
+               MALLOC(newofile, struct file **, nfiles * OFILESIZE,
+                   M_FILEDESC, M_WAITOK);
+               newofileflags = (char *) &newofile[nfiles];
+               /*
+                * Copy the existing ofile and ofileflags arrays
+                * and zero the new portion of each array.
+                */
+               bcopy(fdp->fd_ofiles, newofile,
+                       (i = sizeof(struct file *) * fdp->fd_nfiles));
+               bzero((char *)newofile + i, nfiles * sizeof(struct file *) - i);
+               bcopy(fdp->fd_ofileflags, newofileflags,
+                       (i = sizeof(char) * fdp->fd_nfiles));
+               bzero(newofileflags + i, nfiles * sizeof(char) - i);
+               if (fdp->fd_nfiles > NDFILE)
+                       FREE(fdp->fd_ofiles, M_FILEDESC);
+               fdp->fd_ofiles = newofile;
+               fdp->fd_ofileflags = newofileflags;
+               fdp->fd_nfiles = nfiles;
+               fdexpand++;

        }
        }

-       return (EMFILE);

 }
 
 /*
 }
 
 /*

- * Check to see if any user file descriptors are available.
+ * Check to see whether n user file descriptors
+ * are available to the process p.

  */
  */

-ufavail()
+fdavail(p, n)
+       struct proc *p;
+       register int n;

 {
 {

-       register int i, avail = 0;
-
-       for (i = 0; i < NOFILE; i++)
-               if (u.u_ofile[i] == NULL)
-                       avail++;
-       return (avail);
+       register struct filedesc *fdp = p->p_fd;
+       register struct file **fpp;
+       register int i;
+
+       if ((i = p->p_rlimit[RLIMIT_OFILE].rlim_cur - fdp->fd_nfiles) > 0 &&
+           (n -= i) <= 0)
+               return (1);
+       fpp = &fdp->fd_ofiles[fdp->fd_freefile];
+       for (i = fdp->fd_nfiles - fdp->fd_freefile; --i >= 0; fpp++)
+               if (*fpp == NULL && --n <= 0)
+                       return (1);
+       return (0);

 }
 
 }
 

-struct file *lastf;

 /*
 /*

- * Allocate a user file descriptor
- * and a file structure.
- * Initialize the descriptor
- * to point at the file structure.
+ * Create a new open file structure and allocate
+ * a file decriptor for the process that refers to it.

  */
  */

-falloc(resultfp, resultfd)
+falloc(p, resultfp, resultfd)
+       register struct proc *p;

        struct file **resultfp;
        int *resultfd;
 {
        struct file **resultfp;
        int *resultfd;
 {

-       register struct file *fp;
+       register struct file *fp, *fq, **fpp;

        int error, i;
 
        int error, i;
 

-       if (error = ufalloc(0, &i))
+       if (error = fdalloc(p, 0, &i))

                return (error);
                return (error);

-       if (lastf == 0)
-               lastf = file;
-       for (fp = lastf; fp < fileNFILE; fp++)
-               if (fp->f_count == 0)
-                       goto slot;
-       for (fp = file; fp < lastf; fp++)
-               if (fp->f_count == 0)
-                       goto slot;
-       tablefull("file");
-       return (ENFILE);
-slot:
-       u.u_ofile[i] = fp;
+       if (nfiles >= maxfiles) {
+               tablefull("file");
+               return (ENFILE);
+       }
+       /*
+        * Allocate a new file descriptor.
+        * If the process has file descriptor zero open, add to the list
+        * of open files at that point, otherwise put it at the front of
+        * the list of open files.
+        */
+       nfiles++;
+       MALLOC(fp, struct file *, sizeof(struct file), M_FILE, M_WAITOK);
+       if (fq = p->p_fd->fd_ofiles[0])
+               fpp = &fq->f_filef;
+       else
+               fpp = &filehead;
+       p->p_fd->fd_ofiles[i] = fp;
+       if (fq = *fpp)
+               fq->f_fileb = &fp->f_filef;
+       fp->f_filef = fq;
+       fp->f_fileb = fpp;
+       *fpp = fp;

        fp->f_count = 1;
        fp->f_count = 1;

-       fp->f_data = 0;
+       fp->f_msgcount = 0;

        fp->f_offset = 0;
        fp->f_offset = 0;

-       fp->f_cred = u.u_cred;
+       fp->f_cred = p->p_ucred;

        crhold(fp->f_cred);
        crhold(fp->f_cred);

-       lastf = fp + 1;

        if (resultfp)
                *resultfp = fp;
        if (resultfd)
        if (resultfp)
                *resultfp = fp;
        if (resultfd)


@@ -388,61 +505,196 @@ slot:
        return (0);
 }
 
        return (0);
 }
 

+/*
+ * Free a file descriptor.
+ */
+ffree(fp)
+       register struct file *fp;
+{
+       register struct file *fq;
+
+       if (fq = fp->f_filef)
+               fq->f_fileb = fp->f_fileb;
+       *fp->f_fileb = fq;
+       crfree(fp->f_cred);
+#ifdef DIAGNOSTIC
+       fp->f_filef = NULL;
+       fp->f_fileb = NULL;
+       fp->f_count = 0;
+#endif
+       nfiles--;
+       FREE(fp, M_FILE);
+}
+
+/*
+ * Copy a filedesc structure.
+ */
+struct filedesc *
+fdcopy(p)
+       struct proc *p;
+{
+       register struct filedesc *newfdp, *fdp = p->p_fd;
+       register struct file **fpp;
+       register int i;
+
+       MALLOC(newfdp, struct filedesc *, sizeof(struct filedesc0),
+           M_FILEDESC, M_WAITOK);
+       bcopy(fdp, newfdp, sizeof(struct filedesc));
+       VREF(newfdp->fd_cdir);
+       if (newfdp->fd_rdir)
+               VREF(newfdp->fd_rdir);
+       newfdp->fd_refcnt = 1;
+
+       /*
+        * If the number of open files fits in the internal arrays
+        * of the open file structure, use them, otherwise allocate
+        * additional memory for the number of descriptors currently
+        * in use.
+        */
+       if (newfdp->fd_lastfile < NDFILE) {
+               newfdp->fd_ofiles = ((struct filedesc0 *) newfdp)->fd_dfiles;
+               newfdp->fd_ofileflags =
+                   ((struct filedesc0 *) newfdp)->fd_dfileflags;
+               i = NDFILE;
+       } else {
+               /*
+                * Compute the smallest multiple of NDEXTENT needed
+                * for the file descriptors currently in use,
+                * allowing the table to shrink.
+                */
+               i = newfdp->fd_nfiles;
+               while (i > 2 * NDEXTENT && i >= newfdp->fd_lastfile * 2)
+                       i /= 2;
+               MALLOC(newfdp->fd_ofiles, struct file **, i * OFILESIZE,
+                   M_FILEDESC, M_WAITOK);
+               newfdp->fd_ofileflags = (char *) &newfdp->fd_ofiles[i];
+       }
+       newfdp->fd_nfiles = i;
+       bcopy(fdp->fd_ofiles, newfdp->fd_ofiles, i * sizeof(struct file **));
+       bcopy(fdp->fd_ofileflags, newfdp->fd_ofileflags, i * sizeof(char));
+       fpp = newfdp->fd_ofiles;
+       for (i = newfdp->fd_lastfile; i-- >= 0; fpp++)
+               if (*fpp != NULL)
+                       (*fpp)->f_count++;
+       return (newfdp);
+}
+
+/*
+ * Release a filedesc structure.
+ */
+void
+fdfree(p)
+       struct proc *p;
+{
+       register struct filedesc *fdp = p->p_fd;
+       struct file **fpp;
+       register int i;
+
+       if (--fdp->fd_refcnt > 0)
+               return;
+       fpp = fdp->fd_ofiles;
+       for (i = fdp->fd_lastfile; i-- >= 0; fpp++)
+               if (*fpp)
+                       (void) closef(*fpp, p);
+       if (fdp->fd_nfiles > NDFILE)
+               FREE(fdp->fd_ofiles, M_FILEDESC);
+       vrele(fdp->fd_cdir);
+       if (fdp->fd_rdir)
+               vrele(fdp->fd_rdir);
+       FREE(fdp, M_FILEDESC);
+}
+

 /*
  * Internal form of close.
  * Decrement reference count on file structure.
  */
 /*
  * Internal form of close.
  * Decrement reference count on file structure.
  */

-closef(fp)
+closef(fp, p)

        register struct file *fp;
        register struct file *fp;

+       register struct proc *p;

 {
 {

+       struct vnode *vp;
+       struct flock lf;

        int error;
 
        if (fp == NULL)
                return (0);
        int error;
 
        if (fp == NULL)
                return (0);

-       if (fp->f_count > 1) {
-               fp->f_count--;
+       /*
+        * POSIX record locking dictates that any close releases ALL
+        * locks owned by this process.  This is handled by setting
+        * a flag in the unlock to free ONLY locks obeying POSIX
+        * semantics, and not to free BSD-style file locks.
+        */
+       if ((p->p_flag & SADVLCK) && fp->f_type == DTYPE_VNODE) {
+               lf.l_whence = SEEK_SET;
+               lf.l_start = 0;
+               lf.l_len = 0;
+               lf.l_type = F_UNLCK;
+               vp = (struct vnode *)fp->f_data;
+               (void) VOP_ADVLOCK(vp, (caddr_t)p, F_UNLCK, &lf, F_POSIX);
+       }
+       if (--fp->f_count > 0)

                return (0);
                return (0);

+       if (fp->f_count < 0)
+               panic("closef: count < 0");
+       if ((fp->f_flag & FHASLOCK) && fp->f_type == DTYPE_VNODE) {
+               lf.l_whence = SEEK_SET;
+               lf.l_start = 0;
+               lf.l_len = 0;
+               lf.l_type = F_UNLCK;
+               vp = (struct vnode *)fp->f_data;
+               (void) VOP_ADVLOCK(vp, (caddr_t)fp, F_UNLCK, &lf, F_FLOCK);

        }
        }

-       if (fp->f_count < 1)
-               panic("closef: count < 1");
-       error = (*fp->f_ops->fo_close)(fp);
-       crfree(fp->f_cred);
-       fp->f_count = 0;
+       error = (*fp->f_ops->fo_close)(fp, p);
+       ffree(fp);

        return (error);
 }
 
 /*
  * Apply an advisory lock on a file descriptor.
        return (error);
 }
 
 /*
  * Apply an advisory lock on a file descriptor.

+ *
+ * Just attempt to get a record lock of the requested type on
+ * the entire file (l_whence = SEEK_SET, l_start = 0, l_len = 0).

  */
  */

+

 /* ARGSUSED */
 flock(p, uap, retval)
        struct proc *p;
        register struct args {
 /* ARGSUSED */
 flock(p, uap, retval)
        struct proc *p;
        register struct args {

-               int     fdes;
+               int     fd;

                int     how;
        } *uap;
        int *retval;
 {
                int     how;
        } *uap;
        int *retval;
 {

+       register struct filedesc *fdp = p->p_fd;

        register struct file *fp;
        register struct file *fp;

+       struct vnode *vp;
+       struct flock lf;
+       int error;

 
 

-       if ((unsigned)uap->fdes >= NOFILE ||
-           (fp = u.u_ofile[uap->fdes]) == NULL)
+       if ((unsigned)uap->fd >= fdp->fd_nfiles ||
+           (fp = fdp->fd_ofiles[uap->fd]) == NULL)

                return (EBADF);
        if (fp->f_type != DTYPE_VNODE)
                return (EOPNOTSUPP);
                return (EBADF);
        if (fp->f_type != DTYPE_VNODE)
                return (EOPNOTSUPP);

+       vp = (struct vnode *)fp->f_data;
+       lf.l_whence = SEEK_SET;
+       lf.l_start = 0;
+       lf.l_len = 0;

        if (uap->how & LOCK_UN) {
        if (uap->how & LOCK_UN) {

-               vn_unlock(fp, FSHLOCK|FEXLOCK);
-               return (0);
+               lf.l_type = F_UNLCK;
+               fp->f_flag &= ~FHASLOCK;
+               return (VOP_ADVLOCK(vp, (caddr_t)fp, F_UNLCK, &lf, F_FLOCK));

        }
        }

-       if ((uap->how & (LOCK_SH | LOCK_EX)) == 0)
-               return (0);                             /* error? */

        if (uap->how & LOCK_EX)
        if (uap->how & LOCK_EX)

-               uap->how &= ~LOCK_SH;
-       /* avoid work... */
-       if ((fp->f_flag & FEXLOCK) && (uap->how & LOCK_EX) ||
-           (fp->f_flag & FSHLOCK) && (uap->how & LOCK_SH))
-               return (0);
-       return (vn_lock(fp, uap->how));
+               lf.l_type = F_WRLCK;
+       else if (uap->how & LOCK_SH)
+               lf.l_type = F_RDLCK;
+       else
+               return (EBADF);
+       fp->f_flag |= FHASLOCK;
+       if (uap->how & LOCK_NB)
+               return (VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, F_FLOCK));
+       return (VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, F_FLOCK|F_WAIT));

 }
 
 /*
 }
 
 /*


@@ -458,24 +710,24 @@ fdopen(dev, mode, type)
        dev_t dev;
        int mode, type;
 {
        dev_t dev;
        int mode, type;
 {

-       struct proc *p = u.u_procp;             /* XXX */

 
        /*
 
        /*

-        * XXX Kludge: set p->p_dupfd to contain the value of the
+        * XXX Kludge: set curproc->p_dupfd to contain the value of the

         * the file descriptor being sought for duplication. The error 
         * return ensures that the vnode for this device will be released
         * by vn_open. Open will detect this special error and take the
         * actions in dupfdopen below. Other callers of vn_open or VOP_OPEN
         * will simply report the error.
         */
         * the file descriptor being sought for duplication. The error 
         * return ensures that the vnode for this device will be released
         * by vn_open. Open will detect this special error and take the
         * actions in dupfdopen below. Other callers of vn_open or VOP_OPEN
         * will simply report the error.
         */

-       p->p_dupfd = minor(dev);
+       curproc->p_dupfd = minor(dev);          /* XXX */

        return (ENODEV);
 }
 
 /*
  * Duplicate the specified descriptor to a free descriptor.
  */
        return (ENODEV);
 }
 
 /*
  * Duplicate the specified descriptor to a free descriptor.
  */

-dupfdopen(indx, dfd, mode)
+dupfdopen(fdp, indx, dfd, mode)
+       register struct filedesc *fdp;

        register int indx, dfd;
        int mode;
 {
        register int indx, dfd;
        int mode;
 {


@@ -489,9 +741,9 @@ dupfdopen(indx, dfd, mode)
         * falloc could allocate an already closed to-be-dup'd descriptor
         * as the new descriptor.
         */
         * falloc could allocate an already closed to-be-dup'd descriptor
         * as the new descriptor.
         */

-       fp = u.u_ofile[indx];
-       if ((u_int)dfd >= NOFILE || (wfp = u.u_ofile[dfd]) == NULL ||
-           fp == wfp)
+       fp = fdp->fd_ofiles[indx];
+       if ((u_int)dfd >= fdp->fd_nfiles ||
+           (wfp = fdp->fd_ofiles[dfd]) == NULL || fp == wfp)

                return (EBADF);
 
        /*
                return (EBADF);
 
        /*


@@ -500,10 +752,10 @@ dupfdopen(indx, dfd, mode)
         */
        if (((mode & (FREAD|FWRITE)) | wfp->f_flag) != wfp->f_flag)
                return (EACCES);
         */
        if (((mode & (FREAD|FWRITE)) | wfp->f_flag) != wfp->f_flag)
                return (EACCES);

-       u.u_ofile[indx] = wfp;
-       u.u_pofile[indx] = u.u_pofile[dfd];
+       fdp->fd_ofiles[indx] = wfp;
+       fdp->fd_ofileflags[indx] = fdp->fd_ofileflags[dfd];

        wfp->f_count++;
        wfp->f_count++;

-       if (indx > u.u_lastfile)
-               u.u_lastfile = indx;
+       if (indx > fdp->fd_lastfile)
+               fdp->fd_lastfile = indx;

        return (0);
 }
        return (0);
 }